Biswadip Mandal
Applied Scientist at Amazon
Seattle, Washington, United States
Summary
👤
Senior🎓
Top SchoolBiswadip Mandal is an applied scientist with nine years of experience building production-ready NLP and ML systems, currently focusing on LLM fine-tuning at Amazon in Seattle. He combines a strong academic foundation—MS thesis work on scientific document summarization with publications at NAACL 2022 and LREC-COLING 2024—with hands-on industry impact across fraud prevention, product retrieval and ranking, and clinical NLP. At Staples he progressed from engineer to MLE III, and his open-source contributions include meaningful improvements to the popular AutoKeras AutoML library, adding type hints and enhancing structured-data and transformer/RNN block integration. Comfortable moving between research and engineering, he leverages mathematical approaches to ML and a publication record (Google Scholar profile) to inform practical model design. An IIT Bombay chemical engineering alumnus who pivoted into machine learning, he brings cross-disciplinary problem solving and a track record of shipping models in both startup and large-company settings.
9 years of coding experience
8 years of employment as a software developer
Master of Science - MS, Computer Science (Thesis), Master of Science - MS, Computer Science (Thesis) at The University of Texas at DallasIndian Institute of Technology Bombay
English, Bengali, Hindi
Stackoverflow
544reputation
51kreached
13answers
7questions
Github Skills (15)
keras10
machine-learning10
deeplearning-ai10
deep-learning10
tensorflow10
python10
automl9
neural-architecture-search9
rnn-model8
transformers8
n8
pandas7
confusion-matrix6
word2vec6
numpy6
Programming languages (6)
Github contributions (5)
keras-team/autokeras
Aug 2020 - Mar 2021
AutoML library for deep learning
Role in this project:
ML Engineer
ML EngineerContributions:5 commits, 7 PRs, 13 comments in 7 months
Contributions summary:Biswadip primarily contributed to the AutoKeras library by adding type hints and refining existing code. Their work involved modifying blocks and tasks for structured data, which included adjusting parameters and data handling. The user also contributed to the model's functionality, particularly involving choice of RNN and transformer blocks. Finally, the user contributed to the integration of datasets within the library.
pytorchpythondeep-learningautomated-machine-learningneural-architecture-search
mandalbiswadip/genderPrediction
Sep 2018 - Sep 2021
Gender prediction from Indian names
Contributions:35 commits, 1 PR, 25 pushes in 3 years
genderpredictiongender-prediction
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.